福岡工業大学学術機関リポジトリ

Title

Distribution of Electrically Active Nickel Atoms in

Dislocation-Free N-and P-Type Silicon Crystals Measured by Deep Level

Transient Spectroscopy

Author(s)

田中秀司

Citation

福岡工業大学研究論集　第39巻第1号 　P13-P15

Issue Date

2006-9

URI

http://hdl.handle.net/11478/835

Right

Type

Departmental Bulletin Paper

Textversion publisher

福岡工業大学　機関リポジトリ　

FITREPO

福岡J業大学研究論集 Res. Bull. Fukuoka Inヽt. Tech. → Vol. 39 No. l (2006) 13- 15

-13-Distribution of Electrically Active Nickel Atoms

in Dislocation-Free N- and P-Type Silicon Crystals

Measured by Deep Level Transient Spectroscopy

Shuji TANAKA (Dcpa1imenl ofinformation Electronics, Faculty ofEnginee1ingJ Hajime KITAGAWA (Department of Information Electronics, Faculty of Engineering)

Abstract

Distribution profiles of electrically active nickel atoms in n-and p-typc dislocation-free silicon arc measured by means of the deep level transient spectroscopy (DLTS) on Schottky barrier diodes (SBD) fabricated on nickel-doped silicon. The processes of lapping-off the surface. etching, SBD formation and DLTS measurement were repeated on one sample until the total removed-off thickness exceeded the half of the initial sample thickness. The distribution profiles were evaluated by measuring the con­ centrations of the clcctron trap (nickel acceptor level) in n-type and the hole trap (nickel donor level) in p-type ,ilicon as functions of入I I. where x is the di: ヽtancc from the surface and I is the s皿pie thick­

ness.

The distributions manifest U-shaped diffusion profiles irrespective of one-sided or double-sided diffusion conditions. The experimental results have shown. in the bulk, the flat profiles peculiar to those according to the dissociative mechardsm of diffusion in which the sinks and sourees of lattice va­ cancics are present in the bulk.

Keywordiヽ： nickel distrihution, dislocation-free silicon, n-silicon, p-silicon, U-shaped pro(ile, DLTS

1. Introduction

Nickel is known as one of the fastest diffwヽing elc­ ments in silicon among 3d transition-metal impurities. A large fraction of total nickel atoms dissolved into a silicon C応'Stal stays at interstitial sites and precipitate, in the bulk'" during the heat treatment at high temperature or during quenching. The rest fraction not more than 10·of total contents, which stays at substitutio叫sites'ど'. is elec­ trically ionizahle. or electiically active. Interstitial nickel atoms (Ni,) have been recently reported11 _{lo be predomi­} nantly neutral, or electrically inactive. The substitutional

平成18年5月29 [J受付

nickel atoms (Ni,) in silicon introduceごan acceptor level at Ee·- 0.4 7 e V and a donor level atじ+0.18eV、where£, and£, are energies of the bottom of the conduction band and the top of the valence b出1d, respectively.

Because nickel atoms occupy both interstitial and sub­ stitutional sites, the diffusion of nickel atoms is accompa­ niecl by the site exchange between interstitial and substitu­ tional sites including intrinsic point defects. Not only in clislocatecl silicon6 _{but also in dislocation-free silicon}7_{· も＇} the site exchange mechanism or nickel atoms in silicon has been reported to be due to the dissociative mechanism in which the dominant point defects acting on the site ex­ change are vacancies.

In the present study, thc distribution of the concentra­ tion. C, ol'Nis in dislocation-free n-and p-type silicon

crys-14 Distribution ofElectrically Active Nickel Atoms in Dislocation Free N and P Type Silicon Crystals Measured by Deep Level Transient Spectroscopy(TANAKA·KITAGAWA) tals are measured by deep level transient spectroscopy

(DLTS) on a Schottky barrier diode (SBD). 2. Experimental

The silicon crystals used for the experiment were phosphorus doped n type and floating zoned silicon in both dislocated and dislocation free silicon. Phosphorus content in these crystals was 2.7 X I oi_{: to l.2X 10}14 _cm 3 _in

dislocated silicon and ran°ed from こ 1 1 X 1013 to 1.4 X 1014

cm―3 in dislocation free silicon.

Nickel was evaporated only onto one side of surfaces of silicon slices because it has been already confirmed ex perimentally91 _{that C, shows identical U shaped diffusion}

profile irrespective of one sided or double sided diffusion conditions. The evaporated side is referred to as the front side and the opposite surface is referred to as the rear side.

Nickel diffusion was carried out in flowing nitrogen gaヽambient. The diffusion temperature T0 and mffusion time t0 are T0 = 980℃ and 1ぃ = I 0, 30 and 120 min for n

type silicon, and T0 = 950℃ and t0 = 15, 30 and 60 min for

p type silicon.

After the heat treatment, the slice was lightly lapped and etched to remove nickel silicide and silicon oxide layer from the front side. In the present paper, x = 0 refers to this lapped and etched front side, where x is the distance from the surface. The rear side was never treated after thin layer was once grinded off and etched off.

The DLTS measurements were performed on an SBD fom1ed by means of gold evaporation on n type and alumi̶ num evaporation on p type silicon with a reverse bias of 5 V, a filling pulse bias oro V, a rate window of0.5/5.0 ms and an injection pulse bias of 500µs. The concentration of Cs was evaluated from the peak height of the DLTS associated with the nickel acceptor level in n type and the nickel do nor level in p type silicon, and steady state capacitance.

To obtain the diヽtribution of Cs, the processes of lapping off the surface. etching, SBD rormation and DLTS measurement were repeated on one sample until the total removed off thickness cxccedcd the half of the initial sam ple thickness.

3. Results and Discussion

Typical DLTS signals are shown in Fig. 1 for nickel doped n type silicon and in Fig. 2 for nickel doped p typc silicon. The DLTS peak observed at about 270K is identi fied as the signal due to an electron emission process at a nickel acceptor level (electron trap B5l_{)�The activation en}

crgy of electron emission is Ee 0.4 7 c V which is in good

agreement with the value repo1ted in references 4 and 5. In

ー

。

(dd)J 17 {Ulm!S SL1Q

E OA7 eV

C I I I

75 100

150

200 250

300

T e m p e r a t u r e T ( K )

Fig. 1. DLTS spectrnm of n type silicon doped with nickel with T0 = 980℃ and t0 = 10 min at x I l = 0.086 with l =

0.150cm. CJ; 叫

゜

ヽ

_u

0.. 可

一

,rJ)

乱

('j , ;

1

旦

Control

2 : ;

誓

eV

75 100

150

200 250 300

T e m p e r a t u r e T ( K )

Fig. 2. DLTS spectrum of p type silicon doped with nickel with T0 = 950℃ and t0 = 10 min at x / l = 0.06 I with l =

Distribution of Electrically Aclil'e Nickel Atoms in Dislocation 「rce li and P Type Silicon Cr)温1

、

, !easuredby Deep Level Transient SpcctroヽしOp)爪\AKA·KITAGAWA) 15

.。c,

（）

。

ー (t,

ru:'ln

三

ど）s>

t△

0

。。

_(a)

0 0

□

(b)

ロ ロ □ ロ

△

0.0

△

△

_{△ △△}

(c)

0.2

0.4

0.6

x!l

0.8

1.0

Fig.3.Di叩tribution of Cs inn type silicon. 111 = 120 min (a)、

30 min (b) and 10 min (c) at TD= 980゜C .

ࢎ

ࢎ

。

ヽ

co

. 0

OL

。

ー ([ Ul;)n01 5 x ) こう

(a)

0 O o o O o o

ロ 屯 ロ ロ □ □ ロ

△

(b)

△△ △△ △ △ △

(c)

△

_{△ △}

2

0.0

0.2 0.4

0.6

0.8

x/l

1.0

Fig. 4. Distribution of Cs in p type silicon. Ir, = 60 nlin (a),

30 min (b) and 15 min (c) at Tl)= 950t:.

nickel doped p type silicon, the DLTS peak at about 90K 1s identified as the signal due to a hole emission process at a nickel donor level (hole trap C''). The hole e1nission activa tion energy of the nickel donor level is Ev + 0. 18 e V as

shown in Fig. (2). This value iヽin good agreement with that reported in reference, 4 and 5

The distribution profiles of C, are evaluated by meas uring the concentrations of the electron trap Binn type and the hole trap C in p type silicon as「unction of x I I, where I is the sample thickness. The experimentally obtained dis tributions of Cs are plotted in Fig. 3 for the electron trap B

and in Fig. 4 for hole trap C. Roughlyヽpeaking, the distri

butionヽin both Figs. 3 and 4 manifest U ヽhaped profiles though only the half of the entire profiles are shown. In ad dition, the profiles of the electron trap B and hole trap C are qualitatively alike. Such aヽimilarity is to be expected only when the two trap levels are due to the different charge state of the same nickel atomヽ，0「an amphoteric nickel center.

Reg,trding a theoretical consideration, theoretical cal culation based on the dissociative and kick out mecha niヽms of diffusion is currently in progresヽ. It should be noted that the experimental results show the flat profile in the crystal bulk. Such flat profiles are peculiar to the pro files according to the dissociative mechanism of diffusion in which the sinks and sources of lattice vacancies are pre sent in the hulk.

In summary. the distribution o「substitutional nickel atoms in dislocation free n and p type silicon has been measured by means ofDLTS method. The experimental re sults show the flat profiles peculiar to those according to the dissociative mechanism of diffusion in which the sinks and sources oflattice vacancies are present in the bulk.

References

1) M. Yoshida and K. Furu曲ho: Jpn. J.Appl. Phys. 3 (1964)

521.

2 1 M. Yoshida and K. Saito: .Jpn. J. Appl. Phyヽ. 6 (1967) 573.

3 ) A A Istratov, P. Zhang, R. J. Mcdonald, A. Smith, M. Seacrist, J. Moreland. J. Shen. R. Wahlich and E. R. We ber: The Proceedings of'The 4th International Sympo sium on Advanced Science and Technology of Silicon Materials (JSPS Si S) mposium, Abstract) (2004) 104. 1 I H. Kitagawa and H. Nakashima: Jpn .I. Appl. Phys. 28

(1989) 305

5) H. Kitagawa, S. Tanaka. H. Nakashima and M. Yoshida: J. Electron. Mater. 20 (1991) 441.

6) H. Kitagawa. K. Hashimoto and M. Yoshida: Phyヽ1ca 116B (1983) 323.

, ) S. Tanaka. T. lkari and H. Kitagawa: Jpn. J. Appl. Phys. 40 (2001) 3063.

8) S. Tanaka, H. Kitagawa and T. Ikari: Physica B 308 310 (2001)427.

9) S. Tanaka, T. Ikari and H. Kitagawa: Jpn. J. Appl. Phys. 43 (2004) 7458.